10 things R can do that might surprise you · Simply Statistics
Over the previous couple of weeks I’ve had a few interactions with people from the pc science world who have been fairly disparaging of the R programming language. Plenty of the critism centered on perceived limitations of R to statistical evaluation.
It’s true, R does have a massively complete listing of study packages on CRAN, Bioconductor, Neuroconductor, and ROpenSci in addition to nice bundle administration. As I used to be having these conversations I spotted that R has grown right into a multi-purpose connective language for issues past simply knowledge evaluation. However that the performance isn’t all the time as well-known outdoors of the R neighborhood. So this submit is about a few of the ridiculously superior options of R that will or is probably not as broadly recognized. Listed below are 10 issues R can do you won’t have recognized about, constructing on Kara’s great tweet thread about lighthearted things to do with R.
1. You possibly can write reproducible Phrase or Powerpoint paperwork from R markdown
2. You possibly can construct and host interactive internet apps in only a few strains of code
In only a few strains of code you possibly can create interactive internet apps in R. For instance, in just 36 lines of code you possibly can create an interactive dashboard to discover your BMI in relation to the NHANES pattern utilizing the flexdashboard bundle.
3. You possibly can host your internet apps in yet one more line of R code
The opposite cool factor about constructing internet apps in R is that you would be able to get them up on the net with simply one other line or two of R code utilizing the rsconnect bundle. You possibly can put them up by yourself server or, even simpler, host them on a cloud server like shinyapps.io.
4. You possibly can connect with virtually any database below the solar and pull knowledge with dplyr/dbplyr
It’s very easy to hook up with virtually any database (native or distant) utilizing the dbplyr bundle. This makes it potential for an R consumer to work independently pulling knowledge from almost all common database types. You can even use specialised packages like bigrquery to work immediately with BigQuery and different excessive efficiency knowledge shops.
5. You should use the identical dplyr grammar regionally or on knowledge on a number of totally different knowledge shops
When you learn to do primary knowledge tranforms with dplyr, you possibly can apply the identical code to research knowledge regionally in your pc or remotely on any of the above databases or knowledge shops. This simplifies and unifies knowledge manipulation throughout a number of totally different databases and languages.
6. You possibly can match deep studying fashions with keras and Tensorflow
The keras bundle permits you to match each pre-trained and denovo deep studying fashions immediately from R. You can even work with the direct TensorFlow interface to suit the identical sort of fashions.
7. You possibly can construct APIs and serve them from R
The plumbr R bundle enables you to convert R features to internet APIs that may be built-in into downstream purposes. When you have Rstudio Join you may also deploy them as easily as you deploy internet apps.
8. You can also make online game interfaces with R
Not solely are you able to deploy internet apps, you may make them into superior video video games in R. The nessy bundle enables you to create NES wanting Shiny apps and deploy them identical to you’d every other Shiny app.
9. You possibly can analyze knowledge utilizing Spark clusters proper from R
Need to match huge, gnarly machine studying fashions on large knowledge units? You are able to do that proper from R utilizing the sparklyr bundle. You should use spark in your Desktop or a monster Spark cluster.
10. You possibly can construct and be taught R interactively in R
The swirl bundle is an R bundle that allows you to construct interactive tutorials for R, proper inside R.
That is not at all a complete listing. You can even connect with AWS Polly and write text to speech synthesis software program or construct Shiny apps that respond to voice commands or construct apps that allow you to mix deep studying and accelerometry knowledge to solid Harry Potter spells. The purpose is that R has turn into far more than only a knowledge evaluation language (though its nonetheless good at that!) and being good at R opens the door to a lot of sensible and funky purposes.